[Semi OT] Redundant WAN Connections

Well the obvious alternative would be satellite.

Unfortunately, as it appears you have noticed, any time that you order services through the existing cables into your facility, you will ultimately have Bellsouth involved. Any of the other providers just lease the lines. Times are changing and other provides will now run cables in, if it is financially feasible. This usually means fiber and a fairly expensive service, to help pay for the cost of construction.

There is a way that you possibly can get by without going that route. You can request redundant services through Bellsouth, stipulating that you want the second service to come from a different direction and a different CO. This may turn out to be a little more expensive, especially if it is a service that you pay for by the route mile. But, in the end, it is a reasonable solution. Call your Bellsouth rep and discuss the options for this.

Also, depending on your traffic volume, consider using vpn for backup (since DSL would come from the telco, cable is probably a better choice - I like cable better, anyway). Much less expensive than a secondary circuit. You will most likely see a lesser service when the main connection is down, but the vpn could be enough to keep you going. ISDN is another possible solution. It all depends on what your usage, budget and expectations are.

Just keep in mind that to really be sure that you'd never be down could be incredibly expensive. So, try to figure what the most reasonable cost/benefit plan of back is, and sell your execs on that.

Hope that helps,

Jim

Thanks Jim, That was a great response...

Hi Rod,

Thanks for the reply, I would rather stay away from satellite, if I am worried about a hurricane knocking out service, I would certainly be worried about a rain shower knocking out connectivity.

Chad

Another poster suggested satellite - you could use that as a 2nd service.

it is worth thinking about what you are protecting and relative risk.

1 way to evaluate resilience is to treat it as insurance - for a given surcharge on my bills on comms then i get a reduction in risk and / or increase in capacity, availability and so on.

That also gives you a start on cost benefit analysis - which you are going to need anyway when someone asks how much and why.........

So - 2 diverse feeds protects your comms links (or those sections you choose to pay for resilience)

but from the info in your post they are both going to feed into a single site - maybe you should think about resilience against local problems such as fire, power outages and flooding - typical ways to start are dual comms romms, server resilience and so on.

the logical limit here is 2 sites, or a disaster recovery scheme that lets you migrate to another location during a major problem

and if that is in a different telco footprint, then you may be able to solve your comms resilience issue as well as a side effect - as long as you can move services across easily.

I'd hazard a guess that the rain shower would end before the destroyed CO was rebuilt.

Touche'

Thanks Rod..... :)

You may want to call fpl fibernet and see if they can offer a solution for you. If they do have something, i'm sure it's going to cost an arm and a leg.

Chad

I'am also from south florida, and let me tell you - redundant CO will not help you. We had a redundant OC-3 from 2 Bellsouth CO's. and guess what - both CO's were down during Wilma - BellSouth did not have a enough generators. The best bet would be to put all the production stuff to colo-site (just make sure they have a DIRECT REDUNDANT fibers to other states/countries).

Roman

Three approaches come to mind:

--first, site the servers you use for testing customer facing applications at your business continuity site with current production instances of your applications, as a mirror or your production servers

--second, if you can have your customer facing servers elsewhere, site them in some co-location space out of the path of natural disasters (can be combined with first approach)

--third, if you *have* to have your customer facing servers on site, then get a second, non-phone company, diverse path, perhaps from someone like Level 3. Make sure you understand the diversity, from the building service entrance on.

your faith in direct fibre is touching, but might be misplaced.

modern DWDM kit gives the most distance between PoP and powered devices like amplifiers - but 100 to 150 Km is all that is usually practical.

Despite the way driver specs are given (ie 110 Km drive distance to quote one Cisco GBIC datasheet), the limiting factor is usually loss.

You might get a bit more if you run the fibres slower than 10G, but it wont be a big improvement.

even bleeding edge "tweaked" stuff (ie undersea / lake crossings where you really dont want to put amps underwater if you can avoid it) only gives you around 300 Km or so before you need amplification.

The flip side is that if you are in a disaster big enough that it wipes out power and other facilities for 100 Km in 2 different directions - then you probably have more important things to worry about than whether your users can log in to get email.....

Hi Roman,

I was thinking about that, moving all the production equipment to someplace like the NAP Of the Americas in Miami.

Thanks for the post!

Chad